A robust design method to improve performance reliability based on degradation characteristic

Component performance degradation and parameter drift due to the effects of environmental condition and operating stress causes the performance measures of electronic system to deviate from the initial design specifications. The non-conformance of performance measures over time is referred as performance reliability, or soft failure. This paper presents a new robust design method for electronic system based on component degradation to improve the system performance reliability. Experiment design and sensitivity analysis are employed to determine the critical component parameters that lead to the deviation of system performance measures. Degradation path models of critical components are established by accelerated degradation test. Using the degradation path models, the degradation characteristic of system performance are evaluated using Monte Carlo simulation. Limited by design specifications, the degradation measure distribution and performance reliability are predicted. Then, the normal and tolerance parameters of sensitive components are optimized by robust design to improve performance reliability and time to soft failure. Finally, a practical implementation of the proposed method is validated by optimizing an LED driver, which shows the practicability and efficiency of the method.